Constitutional Choice, Fiscal Federalism, and

International Environmental Agreements

Wolfgang Buchholz1

University of Regensburg and CESifo

Alexander Haupt2

University of Plymouth and CESifo

Wolfgang Peters3

European University Viadrina

February 2011

1University of Regensburg, Department of Economics, 93040 Regensburg, Germany,

E-mail: wolfgang.buchholz@wiwi-uni-r.de.2University of Plymouth, School of Management, Drake Circus, Plymouth, PL4 8AA,

UK, E-mail: alexander.haupt@plymouth.ac.uk.3European University Viadrina, Department of Economics, P.O. Box 1786, D-15207

Frankfurt (Oder), Germany, E-mail: peters@euv-frankfurt-o.de.

Abstract

In this paper, we analyse how the prospect of international negotiations over trans-

boundary pollution shape the countries�constitutional and political decisions that

precede the bargaining process. We show that the countries�dominant strategy at

the constitutional stage is to assign the authority over environmental policy and

international negotiations to a subsidiary government in that region where the pol-

lution has its origin. Although this decentralization of power is usually accompanied

by a federal transfer scheme, there is no �fair�cost sharing within each country. The

negotiating regions�shares of the national damage falls short of the corresponding

shares of costs. This depresses the bargaining incentives of the regional negotiators

and leads to a suboptimal international environmental agreement from the perspec-

tive of the countries as a whole.

JEL-classi�cations: D72, H41, Q28

Keywords: Constitutional design, �scal federalism, transboundary pollution, Inter-

national Environmental Agreement, strategic delegation.

1 Motivation

Most environmental problems are to some extent international or, more precisely,

interregional by nature. Economic activities not only pollute the environment in the

areas where they take place but also in other regions within the national boundaries

and beyond. In other words, they cause externalities on the national level (between

regions of the same country) and on the international level (between regions of

di¤erent countries). Since non-cooperatively acting governments ignore the negative

impact on the environment outside their jurisdictions, they usually end up in a

prisoners� dilemma. The pollution levels are ine¢ ciently high or, coming to the

same, the abatement levels are ine¢ ciently low.

Internalising international externalities is particularly di¢ cult. In the absence

of a coercive supranational power, the countries rely on international environmental

agreements (IEAs). A federal country can approach the necessary negotiations in

two fundamentally di¤erent ways: The power to decide on an environmental policy

and to negotiate abatement levels can be assigned either to the federal government or

to the governments of the polluting regions. This issue is usually already settled at

a constitutional stage at which the political structure and the distribution of power

are codi�ed. While in the centralised system the negotiating politicians are obliged

to their country as a whole, in the decentralised system they only represent the

interests of their regions. Since the two alternative negotiating agents substantially

di¤er in their interests, the constitutional design profoundly a¤ects the country�s po-

sition in the international bargaining process. A country can use this constitutional

delegation of decisions to gain an advantage in international negotiations.1

The constitutional decision to decentralise or centralise the power to negotiate

IEAs seems to be a discrete one. However, delegation can become a quasi-continuous

decision once federal government applies additional instruments to manipulate the

objectives of the regional government. If decentralisation is adopted, the federal

government can accompany this institutional arrangement by �scal transfers. These

conventional political instruments can serve as a means to �ne tune the bargaining

incentives of the polluting regions. Ultimately, rather long-term constitutional de-

signs combined with relatively short-term political measures substantially shape the

outcome of IEAs.

We provide answers to the following two questions: First, what constitution-and-

policy mix is chosen by a non-cooperatively acting country? And second, how do the

resulting strategies a¤ect the e¢ ciency of an IEA? Both issues are analysed in the

1This insight seems to be common knowledge, as, for instance, the introductory comments of

Kockesen and Ok (2004) illustrate.

1

present paper. To this end, we consider a world with two countries, each of them

consisting of two regions. In each country, one of these two regions hosts �dirty�

industries that generate transboundary pollution. Each country �rst decides at the

constitutional stage whether the federal or the regional authorities are entitled to

determine environmental policy and to negotiate an IEA. Afterwards, each federal

government can implement a transfer scheme to cover the abatement costs of its

country�s polluting region through matching grants and/or to charge the �dirty�

region for the �clean�region�s environmental damage. This mixture of subsidies and

taxes provides an incentive for the �dirty�region to act in the best interest of the

entire country.2 Finally, the governments in charge of international negotiations

enter Nash bargaining over binding abatement targets.

The standard theory on �scal federalism would suggest that assigning power to

the federal government is superior, since the bene�ts of an environmental policy

accrues to both regions in each country.3 However, the present paper demonstrates

that this conventional justi�cation for centralisation does not prevail in the current

context, as the federal governments �nd it optimal to delegate bargaining power to

their regional governments. Such a delegation yields a strategic advantage for the

delegating party.

We show that, at the constitutional stage, the two countries delegate the au-

thority to decide on environmental policy and to negotiate an IEA to the regional

governments. This decentralisation of power is usually accompanied by a federal

transfer scheme. If a federal government implements a combination of matching

grants to the �dirty�region and compensation payments to the �clean�region, then

these transfers are such that the polluting region�s share of the abatement costs ex-

ceeds its share of the country�s abatement bene�ts. Since such a scheme ensures that

a polluting region gains from an IEA relatively less than the country as a whole, it

achieves that the interest of the country�s regional negotiators in abatement is lower

than that of the federal government. Thereby, the constitutional and accompany-

ing political decisions strengthen the country�s bargaining position. But, since both

countries face the same incentives to delegate the bargaining power to the regional

2Kaul et al. (2003) show that matching grants are an adequate measure to enforce international

cooperation on �nancing global public goods. Similarly, Guttman (1978), Danziger and Schnytzer

(1991) and, more recently, Boadway et al.(2007) indicate that matching grants can be e¢ ciency

enhancing.3Oates (1972) famous �decentralization theorem�relates the optimal assignment of powers to

di¤erent government layers to the geographical extent of the externalities that arise from local

activities. The theory of �scal federalism goes back to Musgrave (1959), Olson (1969) and Oates

(1972). It makes use of traditional principles, such as the bene�t and ability-to-pay principle, and

the principle of �scal equivalence and equivalent taxation, which is based on the philosophy of

Thomas Hobbes and John Lock.

2

government and to implement �distortionary�transfers, their attempts to be at an

advantage in the negotiations neutralise each other. Moreover, the resulting IEA

sets ine¢ ciently low abatement targets from the perspectives of the countries as a

whole exactly because the federal governments strategically depress the bargaining

incentives of the negotiators.4

Our conclusions are in contrast to the results in Eckert (2003). She also considers

how the distribution of power in a state a¤ects the outcome of an IEA. But she

concludes that the delegation of bargaining power to the regional level might or

might not be an optimal strategy for a country, while we argue that this kind

of delegation is always an equilibrium choice. The ambiguity of her results are

basically caused by a lack of instruments. In her framework, the federal governments

have no matching grants nor other taxes and transfers at their disposal. Without

accompanying taxes and transfers, a country is left only with a discrete delegation

choice between the federal and the regional government. In this case, it might abstain

from delegating the decision to regional negotiators if their interests di¤er too much

from those of the federal authorities. By contrast, if the federal government can �ne

tune the incentives it gives to the regional government by means of a suitable transfer

scheme as in our model, a decentralised constitution is always optimal from the

perspective of a non-cooperatively acting country. So one major di¤erence between

our approach and that of Eckert (2003) is that our framework �scal federalism

enriches the set of instruments. We will come back to this issue below. A further,

technical di¤erence is that we construct a model that enables us to derive all results

analytically, while Eckert (2003) partly relies on simulations.

The present analysis is also particularly related to the literature on delegation

in bargaining processes. Segendor¤ (1998) and Buchholz et al. (2005) analyse

negotiations between two countries over the provision of a pure international public

good and over cuts in transboundary pollution, respectively. These papers stress

the incentives to strategically delegate bargaining power to agents who di¤er from

the delegating �authority�.5 However, they ignore the role of constitutional choices

4Note that the reason for the ine¤ectiveness of an IEA in the present context is very di¤erent

from that in non-cooperative approaches to international negotiations. Barrett (1994), for instance,

shows that IEAs fail to be e¢ cient in most instances because self-enforcing agreements do often not

support a large number of signatory countries. By contrast, in the current framework negotiations

lead to suboptimal abatement levels because the bargaining authority is delegated to regional

governments whose interests di¤er from those of the country as a whole.5The basic idea� that in misrepresenting the own preferences, an individual can raise her payo¤

in a bargaining solution� was developed in the seminal contributions by Crawford and Varian

(1979), Sobel (1981), and Burtraw (1992). In an environmental context, Hoel (1991) argues that a

country which acts as if it were eco-friendlier than it really is reduces the payo¤ it receives in the

bargaining solution. He does, however, not relate this �nding to delegation or other mechanisms

3

in federal systems as an indirect delegation device.

The paper is organised as follows: In section 2 we describe the basic constitution

of a federation, its �scal instruments and the relationship between abatement costs

and environmental damages. A short description of sequential decisions on the

constitutional design, the political choices and the bargaining agreement completes

this section. The outcome of an international environmental agreement is provided

in section 3. We analyse the federal constitution and its �scal instruments in section

4. In this section, we also explore the entire equilibrium. In section 5, we further

examine the role of constitutional and political choices as devices to strategically

delegate the authority to negotiate an IEA. We summarise our results and discuss

some political implications in the concluding section.

2 Regions, Constitutions, and the Environment

In this section, we present our stylised model that enables us to outline the in-

terplay between the constitutional and political choices on the one hand and the

ensuing international environmental negotiations on the other hand. We consider

a two-country world where environmental externalities between the regions within

a country and between the countries coexist. While the federal government can in

principle adopt suitable measures so that the environmental spillovers between the

domestic regions are internalised, there exists no authority that could do the same

on the international stage. The spillovers between regions of di¤erent countries can

only be tackled by means of an IEA. The decision process outlined below models

not only the international negotiations, but also the preceding constitutional and

political choices in the two countries that substantially a¤ect the succeeding IEA.

These preceding decisions encompass constitutional and political choices. In the

constitutional stage, the power to decide on abatement levels and to negotiate on an

IEA is assigned to the federal governments or to the governments of the polluting re-

gions. For brevity, we refer to these two alternatives as centralised and decentralised

system, respectively. In the political stage, the federal government of each country

decides on a intracountry transfer scheme. Afterward, Nash bargaining between the

governments in charge of negotiations over abatement levels takes place.

2.1 Pollution, Abatement, and Transfers

Consider two neighbouring countries 1 and 2, which are symmetric in every respect.

Each of them consists of two regions. To take into account the fact that �dirty�

that enable a country to commit itself to a strategy which is in con�ict with its true objectives.

4

industries are unevenly distributed, in each country the pollution is assumed to be

generated only in one of the two regions. It, however, damages the environment in

all areas of the same country and abroad. Consequently, the bene�ts of abatement

activities that reduce this transboundary pollution are spread over all four regions.

By contrast, the costs of abatement only occur in the polluting regions.

For instance, think of a �shing industry which is located in the countries�coastal

areas. While these regions bear the costs of tougher regulations, the whole popu-

lation in the two countries bene�ts from the protection of species and the resulting

maintenance of biodiversity. Further examples are the carbon mining industry and

the nuclear power stations along the upper course of a river. Acid waste water and

the waste heat can be only avoided in the upstream areas where the mines and

cold-storage plants are located. The bene�ts of abatement, however, also arise in

the regions downstream.

Denote by ai � 0 the abatement costs of country i�s the polluting region. Theabatement activities diminish pollution and thus environmental damage. An en-

vironmental damage function Di = D(ai + saj), with D0 (�) < 0 and D (�) > 0,

captures the negative relationship between country i�s environmental damage on

the one hand and the abatement costs of country i�s and j�s polluting regions on the

other hand. This damage function is twice-continuously di¤erentiable and strictly

convex. It reaches its maximum Dmax = D(0) if neither region abates. The dam-

age function also ful�ls the Inada-type conditions limai+saj!0D0 (�) = �1 and

limai+saj!1D0 (�) = 0. The parameter s denotes the international spillover. We

assume that s 2 (0; 1) holds, i.e., that abatement in country i�s polluting regioncontributes more to a cleaner environment in country i than in country j.6

Environmental damage Di is measured in monetary terms, and so is thus abate-

ment bene�t Dmax � D(ai + saj). As the abatement bene�t varies only with en-vironmental damage, maximising the overall net bene�t of country i is equivalent

to minimising the total costs, consisting of environmental damage and abatement

costs:

TCFi = D (ai + saj) + ai. (1)

The superscript �F�stands for federal government, which represents the country as

a whole.

Since only a part of the population resides in the polluting region, this region

bears only a part of the country�s environmental damage Di. Denoting this share

by � 2 (0; 1), the damage within the polluting region is equal to �Di. Moreover,

6Note that the upper bound of the interval (0; 1) covers the case of the environment as a pure

international public good, while for the lower bound means that pollution causes environmental

damage only domestically.

5

a region might receive, or pay for, interregional and intracountry transfers, which

are a means of the federal government to shape the bargaining incentives of the pol-

luting region under decentralisation. The federal government has two instruments

at its disposal. Firstly, the �dirty� region can be compensated for its abatement

costs through a matching grant miai. Secondly, the �dirty� region can be forced

to compensate the �clean�region for its environmental damage through a transfer

�iDi. Then, � < 1��, �i = 1��, and � > 1�� represents partial compensation,full compensation, and overcompensation. Both instruments impose some degree

of burden sharing between the regions. However, while the former leaves the prop-

erty right with the �dirty�region, the latter is more in the spirit of a polluter-pays

principle.

To sum up, the polluting region might receive a matching grant miai, which

covers part of the regional abatement costs, and it might be forced to make a com-

pensation payment �iDi. In any case, (mi; �i) � 0 is assumed to hold. Taking thesetransfers into account, the total cost of the polluting region is given by

TCRi = (�+ �i)D (ai + saj) + (1�mi) ai, (2)

where �R�stands for that region�s government, which represents the local population.

2.2 Constitutional Design, Political Choices, Bargaining

A three-stage decision process models not only the international negotiations, but

also the preceding constitutional and political choices that shape the outcome of

the IEA. In the �rst stage (constitutional stage), the federal governments of the two

countries non-cooperatively decide whether the power to determine the abatement

levels in the polluting regions and the power to negotiate over an IEA stay with them

or lie with the governments of the polluting regions. We refer to these two distinct

arrangements as centralised and decentralised systems, respectively.7 Making this

constitutional decision, each federal government takes the strategy of its opponent

as given and chooses the solution that maximizes its payo¤.

If the decentralised system is adopted in a country, the government of the pol-

luting region irrevocably decides on the abatement activities. In this case, only the

regional government can sign a binding IEA. The federal government, however, can

still in�uence the bargaining incentives of the regional authority by implementing a

transfer scheme. As described above, the federal government has a matching grant

7These constitutional choices cannot be revoked, at least not within the relevant time horizon.

So they imply a perfect commitment of the country. In this respect, our approach di¤ers from

models in which principles can revoke their partial commitments at some costs later on. See, for

instance, Muthoo (1996).

6

mi and compensation rate �i at its disposal. Thus, in the second stage (political

stage) it decides on the rates mi and �i.

After the decision on the constitutional and political bargaining framework have

been made in the two countries, the international negotiations can start. In the

third stage (bargaining stage), the two governments in charge of environmental ne-

gotiations bargain over the abatement levels and potential side payments between

them. The outcome of these negotiations is described by the Nash bargaining so-

lution. This solution crucially depends on the threat point, i.e., the outcome that

would be realized if negotiations failed. In that case, the governments which are

responsible for environmental policy set the abatement levels simultaneously and

non-cooperatively. The resulting solution of this Nash game serves as a threat point

for the negotiations.

The following table summarises the sequence of decisions:

Stages Decisions

1. Constitutional stage Governments in charge of negotiations, F or R

2. Political stage Transfer scheme mi and �i3. Bargaining stage IEA: abatement levels a1 and a2 (and side payments)

2.3 Centralization versus Decentralization

A closer look at the total costs (1) and (2) shows that for a compensation rate

�i = 1�� which is not accompanied by a matching grant, i.e. mi = 0, the objective

functions of the federal and the regional governments coincide. In this case, the

regional government has the same bargaining incentive as the federal government.

In fact, the regional government would act as if it were the federal government.

Consequently, the Nash bargaining solution for the decentralized system in the case

of the transfer scheme (mi; �i) = (0; 1��) replicates the centralised system. In thissense, the centralised solution can be considered as a special case of the decentralised

one.

This relationship between the two constitutional designs simpli�es the following

analysis considerably. Conducting the usual backward induction, we do not need

to distinguish between the centralised and decentralised system explicitly. Instead,

we can simply analyse the second and third stage of the game for the case of a

decentralised solution. If either mi 6= 0 or �i 6= 1�� results in the second stage, thedecentralised system is indeed the equilibrium choice. Then, the optimal negotiator

should have bargaining incentives that di¤er from those of the federal government.

This can only be achieved by means of decentralisation. By contrast, if the outcome

is mi = 0 and �i = 1 � � in the political stage, then the federal government

7

would be the optimal negotiator for the country. In this case, there is no need for

decentralisation.

3 International Environmental Agreements

Applying backward induction, we �rst turn to the bargaining solution in the third

stage. Since the non-cooperative abatement levels, which are implemented if the

negotiations fail, determine the threat point, they have to be solved for �rst. Af-

terwards, we provide the Nash-bargaining solution and show how this outcome is

related to countries�transfer schemes. These results are the prerequisites for our

analysis of the constitutional and political decisions in section 4.

3.1 Threat Point

Let us explore the strategies of the governments in charge of environmental policy

when abatement levels are determined non-cooperatively. In this case, the two

governments carry out their optimal policy, taking the strategy of the rival as given

and ignoring the externalities of their policies. Country i�s responsible government

chooses the level ain that minimises its total costs

TCRin = (�+ �i)D (ain + sajn) + (1�mi) ain; (3)

where the additional subscript n indicates that the decisions are made non-cooper-

atively. We obtain the �rst-order condition8

� (�+ �i)D0in = 1�mi for i = 1; 2: (4)

It says that from the perspective of country i�s government in charge of environ-

mental policy, the marginal decline in damage (including compensation payments)

equals marginal abatement costs. If the responsibility for the abatement levels lies

with the federal government (i.e. if mi = 0 and �i = 1 � �), these �gures coincidewith the marginal bene�ts and costs of the country as a whole. By contrast, if

the government of the polluting region is in charge, condition (4) only captures the

marginal bene�ts and costs that accrue to this region.

The �rst-order condition (4) implicitly determines the reaction curves and the

unique Nash equilibrium, which is characterised by an active abatement policy in

both countries if the transfer schemes in the two countries are not too di¤erent.9 The8D0

in denotes the value of the derivative of region i�s damage function with respect to ain+sajnin case of non-cooperatively acting governments.

9In any case, the equilibrium is unique (cf. Buchholz et al., 2005). Moreover, in section 4 we

argue why focusing on an interior solution makes indeed sense.

8

threat point (a1n; a2n) and thus the bargaining solution obviously depend on the two

countries�matching grants m = (m1;m2) and the compensation rates � = (�1; �2).

Since these connections o¤er an opportunity for the federal governments to in�uence

the outcome of an IEA, we are interested in the comparative statics of these variables.

Lemma 1 Threat Point and Transfer Scheme.If the federal government of country i raises the matching grant to, or the compen-

sation payments of, the �dirty�region, then domestic abatement activities increase.

This increase is at least partially o¤set by lower abatement activities abroad. Overall,

environmental damage is diminished in country i and remains unchanged in country

j.

Proof. Using �rst-order condition (4), we obtain

@(ain + sajn)

@mi

>@(ajn + sain)

@mi

= 0 and@(ain + sajn)

@�i>@(ajn + sain)

@�i= 0. (5)

This outcome can only result if @ajn=@mi = �s (@ain=@mi) < 0 and @ajn=@�i =

�s (@ain=@�i) < 0. It directly implies @Din=@mi < 0, @Din=@�i < 0, and @Djn=@mi =

@Djn=@�i = 0.�

The intuition for Proposition 1 is straightforward. A higher matching grant or

compensation rate raises the marginal damage (including compensation payments)

of i�s responsible government relative to its abatement costs. Abatement activities

become more attractive, and the abatement level ain increases. As a consequence,

country j�s marginal damage declines, and thus its incentive to abate. The de-

crease in abatement ajn partially o¤set the increase in ain. In other words, the

countries�choice variables are strategic substitutes like in the standard models of

non-cooperative private provision of public goods. Still, environmental damage di-

minishes in country i because the reduction in domestic pollution more than com-

pensates for the rise of pollution abroad. By contrast, damage is unchanged in

country j.10

3.2 Bargaining over Abatement and Side payments

The equilibrium of the previous section provides the countries�abatement levels if

environmental policies are decided non-cooperatively. This solution determines the

threat point of the negotiations over an IEA. To give the outcome of these negotia-

tions, we apply the Nash-bargaining concept. We also take account of the fact that

agreements on environmental issues are frequently accompanied by (implicit) side

10This conclusion follows from a quasi-linear objective function.

9

payments between countries.11 Thus, any negotiations contain two issues, abatement

levels and potential side payments.

The calculation of the Nash-bargaining solution with side payments is straightfor-

ward. The negotiating governments �rst choose the abatement levels that minimise

their aggregate costs. (Recall that abatement costs as well as the environmental

damages are measured in monetary units.) Afterwards, a side payment from one

government to the other is determined such that the two negotiating parties share

the gains from cooperation evenly.12

The resulting abatement levels a1c and a2c are e¢ cient from the perspective of

the negotiating governments. They minimise their aggregate costs

AC =2Xi=1

[(�+ �i)D (aic + sajc) + (1�mi) aic] , (6)

where the subscript c indicates the cooperative case.

Since this aggregate payo¤ is strictly convex, a unique solution to the allocation

problem (6) exists. The optimal abatement levels are implicitly given by the �rst-

order conditions

� (�+ �i)D0ic � s

��+ �j

�D0jc = 1�mi for i; j = 1; 2 (7)

The left-hand side of condition (7) captures the aggregate marginal decline in damage

(including compensation payments) caused by abatement in either country i or j.

The right-hand side gives the corresponding marginal abatement costs. Trivially,

marginal abatement costs are equal to the generated aggregate marginal cut in

damage (including compensation payments) of the two negotiating parties.

As condition (7) holds for both i = 1 and i = 2, the optimal solution is alterna-

tively characterised by

� (�+ �i)D0ic =

1�mi � s(1�mj)

1� s2 for i; j = 1; 2 : (8)

11The Kyoto Protocol on the emission of substances that cause global warming provides an

example for the incorporation of international transfers into an IEA. We think it is fair to say that

implicit side payments are granted through �generous�grandfathering of certi�cates to Russia and

other countries. Selling these certi�cates on the world market generates transfer-like revenues for

these regions. The role of such side payments is already discussed in Chichilnisky and Heal (1994)

and Eyckmans, Prost and Schokkaert (1993).12Using Nash bargaining with appropriate side payments simpli�es the analysis signi�cantly.

Similar results as in the current framework can be obtained when the modelling approach to

negotiations applies the outside option principle. Then, threat points do not play any role in case

of symmetric countries. Such a kind of modelling reduces the complexity of the economic structure

in a similar way as side payments do.

10

In contrast to its counterpart in the non-cooperative case (4), the spillover parameter

s explicitly appears in condition (8), since the impact of a country�s abatement

activities on the other negotiating party is now taken into account.

If at least one regional government is involved in the negotiations, the resulting

abatement levels depend on the transfers. As a prerequisite of exploring the optimal

schemes from the perspective of a federal government and the question of whether

a federal government decentralises power at all, we analyse the impact of these

schemes on the IEA in more detail.

Lemma 2 Bargaining Outcome and Transfer Scheme.In the cooperative solution, both a higher compensation rate and a higher matching

grant in country i shift the abatement activities from country j to country i. Thus,

the total abatement costs of country i increase. The environmental damage in coun-

try i declines in either case. However, matching grants increase the environmental

damage in country j, while a change in � will leave it unchanged.

Proof. Using �rst-order condition (8), we obtain

@(aic + sajc)

@mi

> 0 >@(ajc + saic)

@mi

and (9 a)

@(aic + sajc)

@�i=

�1� s2

� @aic@�i

>@(ajc + saic)

@�i= 0; respectively. (9 b)

This outcome can only result if the abatement level aic is raised and the level ajcis lowered. It directly implies @Dic=@mi < 0, @Dic=@�i < 0, @Djc=@mi > 0, and

@Djc=@�i = 0.�

As Propositions 1 and 2 show, the impacts of higher matching grants and com-

pensation payments in the cooperative case are very similar to those in the non-

cooperative case. In both cases, such policy changes shift the costs of abatement

from country j to country i.

This broad picture remains valid if side payments are included, as a closer look

on them shows. Since the gains from cooperation are split equally between the two

negotiating parties, side payments Si from i�s government to its opponent in the

negotiations are given by

Si =1

2

8<:[(�+ �i) � (Din �Dic) + (1�mi) � (ain � aic)]| {z }gains of i�s negotiating government from cooperation

(10)

��(�+ �j) � (Djn �Djc) + (1�mj) � (ajn � ajc)

�| {z }gains of j�s negotiating government from cooperation

9>=>;11

where Dic and Din denote country i�s damages in the cooperative solution and in the

non-cooperative threat point, respectively. In the current model, the transfer paid

by one of the negotiating governments equals the payments received by the other,

i.e. Si = �Sj.In the cooperative solution, both a higher compensation rate and a higher match-

ing grant in country i increases the abatement costs including side payments in

country i. As stated above, a rise in country i�s compensation rate or matching

grant increases its abatement costs. However, country i might be rewarded for its

intensi�ed abatement activities by lower side payments Si. In any case, abatement

costs including side payments, i.e. ai + Si, will go up, too.13

4 Constitutional Choice and Transfer Schemes

If the federal government chooses the decentralised solution in the �rst stage, it

still can implement a matching grant and a compensation rate in the second stage.

The transfer schemes are used to �ne tune the bargaining incentive for the country�s

regional government so that the regional negotiators optimally pursue the goal of the

country as a whole. Making its choice on the rates mi and �i, i�s non-cooperatively

acting federal government takes the other country�s negotiating party and, if this is

the regional government too, the transfer scheme abroad as given.

No matter whether the other side�s negotiating party is the federal or the regional

government of that country, its representatives are completely described by the

rates �j and mj. As already explored above, centralisation can only be optimal

if �i = 1 � � and mi = 0 result. Otherwise, decentralisation is country i�s best

strategy.

When choosing the optimal transfer scheme for region i, the federal government

considers environmental damage and abatement costs for the country as a whole

including potential side payments in connection with an IEA. Thus, it minimises

P Fi = TCFi + Si = D (aic(�;m) + sajc(�;m)) + aic(�;m) + Si(�;m), (11)

where � =��i; �j

�and m =(mi;mj).

The federal government indirectly a¤ects the �nal abatement outcome (aic; ajc)

and the side payments Si via its choice of the transfer rates mi and �i. Analysing

the optimal scheme, we �rst focus on interior solutions, but discuss the possibility of

boundary solutions later. Moreover, we concentrate on symmetric equilibria, since

the two countries are identical in every respect.

13An earlier version of this paper discusses this issue in detail and can be obtained upon request.

12

In the case of an interior solution, the optimal transfer scheme is characterized

by the �rst-order conditions

�D0ic

@ (aic + sajc)

@�i=

@ (Si + aic)

@�i(12 a)

�D0ic

@ (aic + sajc)

@mi

=@ (Si + aic)

@mi

. (12 b)

In optimum, the transfer scheme balances two opposing e¤ects already indicated

by the comparative statics in the previous section. On the one hand, both a higher

matching grant and a higher compensation rate indeed increase the global abatement

level aic + sajc and, therefore, decreases country i�s environmental damage (LHSs),

since they provide additional incentives for domestic pollution cuts. On the other

hand, the sum of side payments and abatement costs Si+aic increase (RHSs), since

the grants weaken the country�s bargaining power.

Using the de�nition of the side payments (10), the �rst-order condition (8), and

the results of the comparative statics (5) and (9 b), condition (12 a) yields

@aic@�i

�1�mi � s (1�mj)

�+ �i� [1� s(1�mj)]

�=

(13)1

2

�Din �Dic + s [1�mi + s (1�mj)]

@ain@�i

�> 0:

As the environmental is indeed larger in the non-cooperative solution than in the

cooperative outcome, i.e., Din � Dic > 0, the RHS of (13) is positive.14 Then, the

LHS is also positive in equilibrium. This implies that the inequalities � + �i <[1�mi�s(1�mj)]

[1�s(1�mj)]� 1 are satis�ed. Since �i < 1 � � results, the compensation rate

only partially internalises the externalities between the domestic regions. To put

it di¤erently, the federal government prefers a negotiator that never bears the full

domestic damage. As an immediate consequence of this conclusion, choosing a

decentralised system is a dominant strategy in the constitutional stage. It can never

be optimal that the power to negotiate an IEA lies with the federal government.

To fully assess the implications of this result, we establish the relationship be-

tween the compensation rate and the matching grant in equilibrium. Using (10), the

�rst-order condition (8), and the comparative statics (5) and (9 a), we can rearrange

14To see this, note that the following relationships are implied by the convexity of

the damage function and the conditions (4) and (8): Dic < Din , D0ic > D0

in ,[1�mi + s (1�mj)] =

�1� s2

�< 1 � mi , [1�mj + s (1�mi)] > 0. The last inequality is

ful�lled as long as an interior solution exists and thus the RHS of (8) is positive.

13

the �rst-order condition (12 b):

@aic@mi

�1�mi � s (1�mj)

(1� s2)(�+ �i)+s [1�mi � s(1�mj)]

1� s2 � 1�=

(14)1

2

�aic � ain + s [1�mi + s (1�mj)]

@ain@mi

�+ s

@ajc@mi

[1� �� �i]D0ic > 0.

Since � + �i < 1 follows from (13), the RHS of (14) is positive in a symmetric

equilibrium.15 Thus, the �rst-order condition can be only ful�lled if the LHS is also

positive, which means that the inequality 1�mi > �+�i has to hold. Although the

matching grant covers a part of the abatement costs, the polluting region�s share

of the domestic abatement costs 1 � mi is larger than its share of the domestic

damage � + �i. In this sense, there is no �fair� cost sharing, i.e., the �nancial

burdens associated with the pollution cuts are more than proportionally born by

the polluting region.

In this way, each country delegates bargaining power to a negotiator whose net

gains from a cleaner environment are lower than that of the federal government. This

strategy depresses the willingness of the negotiating regional government to accept

�nancially painful domestic abatement levels. Since it strengthens the country�s bar-

gaining power, the costs of an environmental improvement upon the non-cooperative

solution are shifted to the other country.

This cost-shifting e¤ect explains why countries have an incentive to put regional

governments in charge of environmental policy and negotiations. But why should,

from a country�s perspective, this constitutional choice be accompanied by a trans-

fer scheme? The answer to this question is straightforward. Regional governments

might care �too little�for the environment. Note that the bargaining solution op-

timises, by de�nition, the aggregate payo¤ of the two negotiating regional gov-

ernments. Since they underestimate the �true�payo¤ gains from abatement, the

pollution cuts agreed on in the IEA are suboptimally low from the perspective of

the country as a whole. This negative lack-of-e¤ectiveness dominates the positive

cost-shifting e¤ect if the payo¤ function of the polluting region di¤ers too much

from that of the federal government. To balance these two opposing e¤ects, the re-

gional interests have to be su¢ ciently alined with the federal ones. This goal can be

achieved by means of a transfer scheme, which �ne tunes the bargaining incentives

of the regional government.

Figure 1 a) and b) illustrate this result. The graphs show the relationship be-

tween the federal government�s total costs P Fi and polluting region�s e¤ective damage

15Note that, in the symmetric case, D0ic > D

0in directly implies that gic > gin for i = 1; 2. Thus,

the terms in the �rst bracket of RHS of (14) are positive.

14

PiF

a+bi*

PiF

aa+bi

1 1

a) b)

a +bia

a

Figure 1: Optimal Compensation Rate

share � + �i for a given policy abroad. (For simplicity, assume that the rate mi is

�xed at zero.) In �gure 1 a), the minimum will be reached if rate ��i is implemented.

By contrast, a �pure�decentralisation, i.e., decentralisation without an accompany-

ing transfer scheme, is optimal from the country�s perspective in the case drawn in

�gure 1 b). In this situation, a boundary solution results, since the payo¤ function

of the regional government is already too similar to that of the federal government.

If a negative compensation rate is excluded, the best the federal government can

do is to set the rate �i equal to zero.16 A similar line of reasoning holds for the

matching grant.17 Thus, an interior solution requires a su¢ ciently small parameter

�. To sum up:

Proposition 1 Constitutional Choices and Transfer Schemes.i) In the constitutional stage, each federal government�s dominant strategy is to

put the government of its polluting region in charge of environmental policy and

16Formally, this means that the objective function PFi is already decreasing with �i at �i = 0.17While we cannot exclude a �lower boundary solution�mi = 0 or �i = 0, we can neglect �upper

boundary solutions�. Our preceding analysis already indicates that the outcome mi = 1 cannot be

an equilibrium. In this case, the regional government would not bear any abatement costs, and it

would thus abate as much as possible (see �rst-order conditions (7)). Then, environmental damage

becomes zero at in�nite abatement costs. Such a strategy cannot be optimal even if parts of these

costs were o¤set by side payments. For similar reasons, �too�large rates �i can be also ignored.

Since the solution has to contain rather �moderate�schemes, the relevant intervals of the rates are

reasonably assumed to generate interior solutions (aic; ajc) > 0 and (ain; ajn) > 0 in the third

stage so that we can apply the usual comparative statics.

15

negotiations.

ii) If strictly positive transfers are implemented, the scheme has two properties.

Firstly, the compensation rate internalises the domestic abatement externalities only

partially, i.e., � + ��i < 1. Secondly, in a symmetric equilibrium the polluting

region�s share of abatement costs exceeds its share of the environmental damage,

i.e., 1�m� > � + �� results.

Both countries delegate the power to negotiate an IEA to regional governments

without providing su¢ cient incentives for them to internalise the international ex-

ternalities. Thereby, the federal governments also relinquish the internalisation of

the domestic externalities in order to gain a strategic advantage in the international

negotiations. But the attempts of both countries to shift the costs of an IEA to

the other side neutralise each other. The federal governments end up in a prisoners�

dilemma. In a symmetric equilibrium, the only lasting e¤ect is that the abatement

levels agreed on are suboptimally low from the perspective of the two countries as a

whole. The IEA is ine¢ cient, although it minimises the aggregate costs of the ne-

gotiators, since the countries�free-riding behaviour is simply shifted from the level

of the negotiations to that of the preceding constitutional and political decisions.

Proposition 2 Ine¢ ciency of the IEA.In a symmetric equilibrium, the abatement activities are too low from the perspective

of the countries as a whole. The countries�aggregate payo¤ would be higher if the

two federal governments were in charge of environmental policy and negotiations.

Proof. In a symmetric equilibrium, the two countries make identical choices. Hence,

�D0(a�c(1 + s)) =(1�m�)

(�+ ��)(1 + s)>

1

(1 + s)= �D0(aF (1 + s)); (15)

where aF denotes each country�s abatement level in case in which the federal gov-

ernments negotiate with each other and a�c refers to the abatement level in the

subgame-perfect equilibrium. Note that aF is also the symmetric abatement level

that minimises each country�s total costs D (aF (1 + s)) + aF . (Side payments do

not appear in the payo¤ because they are zero in a symmetric solution.) Since the

damage function D(�) is strictly convex, D0(a�c(1 + s)) < D0(aF (1 + s)) < 0 directly

implies the inequality a�c < aF and thus PF (aF ) < P

F (a�c).�

5 Constitution and Strategic Delegation

The conclusion of the previous section is obvious. Decentralisation is a dominant but

ultimately harmful strategy for both countries. This outcome deviates from Eckert�s

16

(2003) �ndings. In her setting, decentralisation might, but need not necessarily be

the best strategy of a country.

The di¤erence between her and our framework is that she only considers the

constitutional choice and ignores accompanying political instruments of �scal feder-

alism. If the set of possible instruments is constrained in that way, the constitutional

decision is not a priori clear. As explored above, the lack-of-e¤ectiveness can out-

weigh the cost-shifting e¤ect in the case of �pure�decentralisation. This situation

is illustrated in �gure 1a). As long as the share � falls short of the threshold �,

the total costs are lower in the centralised solution than the �purely�decentralised

solution, i.e. P Fi (1) < P Fi (�). Left with the two extreme solutions as the only

choices, the federal government prefers to centralise political power, and in the case

of a symmetric equilibrium the IEA establishes the countries��rst-best solution. By

contrast, if the government has su¢ cient instruments at its disposal, decentralisa-

tion emerges as dominant strategy, and environmental agreements are suboptimal

from the perspective of the federal governments. In this sense, our conclusion is

much stronger and more worrying than that in Eckert (2003).

It is, however, in line with two other contributions on negotiations over the pro-

vision of public goods in general or environmental quality in particular. Segendor¤

(1998) and Buchholz et al. (2005) show that a country�s �authority� gains from

delegating the bargaining power to an �agent�who pays less attention to the envi-

ronment than the �authority�itself.18 In the public-choice approach in Buchholz et

al. (2005), the electorate as a country�s ultimate authority even supports politicians

that attach no weight at all to the environment in the case of a global pollutant.

As a consequence, even international cooperation does not yield any e¤ective im-

provement upon the non-cooperative outcome. The implications are less gloomy in

the present paper. Since negative grants are excluded, the delegation race to the

bottom is bounded from below. Nevertheless, the fundamental tendencies are the

same.

A basic di¤erence between Segendor¤ (1998) and Buchholz et al. (2005) on the

one hand and Eckert (2003) and the present paper on the other hand is that in

18In contrast to these two papers, Siqueira (2003) argues that the preferences of the delegated

agents are consistent with that of the principals if countries cooperatively determine their policy

in the context of international externalities. This di¤erence partly arises because Siqueira (2003)

uses a very di¤erent notion of cooperation. He focuses on an economic and political integration,

i.e., the two governments become a single decision unit. By contrast, Segendor¤�s, Eckert�s and

our solutions are based on Nash bargaining, which encompasses more than maximizing aggregate

payo¤. (See Muthoo, 1999, for an excellent treatment of this approach.) This concept pays, for

instance, particular attention to the threat point. This point, to which the countries fall back if

negotiations fail, does not play a role in the cooperative solution of Siqueira (2003).

17

the former papers there is a continuous delegation choice while in the latter there

is only a discrete one - either the federal government or the government of the

polluting region is entitled to negotiate over an IEA. We show, however, that if the

constitutional decision is accompanied by proper transfers to �ne tune the regional

bargaining incentives, the discrete delegation choice is de facto transformed into a

continuous choice. In this sense, the current paper bridges the gap between the two

strands of literature.

6 Concluding Remarks

In the present paper, we analyse how the prospect of international negotiations over

transboundary pollution shapes the countries�constitutional and political decisions.

We show that, at the constitutional stage, countries have an incentive to assign the

authority over environmental policy and international negotiations to regional gov-

ernments. Even if this decentralisation of power is accompanied by a federal transfer

scheme, the negotiating regions�share of the abatement costs exceed their share of

the environmental damage. The resulting unfair cost sharing within a country means

that the regional negotiators gain less from an IEA than the federal governments.

Depressing the bargaining incentives of the domestic negotiators in this way, the fed-

eral governments intend to shift the abatement costs to the neighbouring country.

Since both countries pursue the same strategy, they neutralise each other, at least

in the symmetric case. While no side can gain an advantage in the negotiations,

the drawback of the non-cooperative constitutional and political decisions becomes

apparent. The abatement levels are suboptimally low from the perspective of the

countries�as a whole, although an IEA is reached. Both countries are worse o¤ than

in the case in which the federal governments enter negotiations.

Our analysis demonstrates that international environmental negotiations by itself

are not su¢ cient to overcome the de�ciencies of non-cooperative behaviour, since

these talks simply shift the problems from the level of the abatement decisions

to the level of the preceding political and constitutional choices. Obviously, the

bargaining framework matters, and a truly cooperative solution has to encompass

the constitutional and political conditions under which negotiations take place.

Unfortunately, not only decisions on intergovernmental transfers and the dis-

tribution of power a¤ect the outcome of IEAs.19 So even if the countries agree on

negotiating at the federal level, this does not guarantee a more e¤ective IEA. Rather

19As shown in Copeland (1990), Buchholz and Konrad (1994) and Buchholz and Haslbeck (1997),

investment decisions and the choice of environmental technologies can be also used to gain an

advantage in ensuing negotiations.

18

it might shift the e¤orts to strategically manipulate the bargaining framework to

other areas. Since complete contracts that precisely regulate all the potentially im-

portant circumstances are not possible, there always remains a refuge for strategic

behaviour prior to negotiations. Therefore, it arises the question of whether limiting

this strategic behaviour partially, even if it is possible, really improves the overall

outcome of IEAs. In any case, since the institutional environment in which inter-

national negotiations are embedded is of fundamental importance for the resulting

agreement, it deserves more attention in future research.

References

Barrett, S. (1994), Self-Enforcing International Environmental Agreements, Oxford

Economic Papers 46, 878-894.

Boadway, R., Z. Song and J.-F. Tremblay (2007), Commitment and Matching

Contributions to Public Goods, discussion paper Queen�s University.

Buchholz, W., and K.A. Konrad (1994), Global Environmental Problems and the

Strategic Choice of Technology, Journal of Economics 60, 299-321.

Buchholz, W., and C. Haslbeck (1997), Strategic Manipulation of Property Rights

in Coasean Bargaining, Journal of Institutional and Theoretical Economics

153, 630-640.

Buchholz, W., A. Haupt and W. Peters (2005), International environmental Agree-

ments and Strategic Voting, Scandinavian Journal of Economics 107, 175-195.

Burtraw, D. (1992), Strategic Delegation in Bargaining, Economics Letters 38,

181-185.

Chichilnisky, G., and G. Heal (1994), Who Should Abate Carbon Emissions? An

International Viewpoint, Economic Letters 44, 443-449.

Copeland, B.R. (1990), Strategic Enhancement and Destruction of Fisheries and

the Environment in the Presence of International Externalities, Journal of

Environmental Economics and Management 19, 212-226.

Crawford, V.P., and H.R. Varian (1979), Distortions of Preferences and the Nash

Theory of Bargaining, Economic Letters 3, 203-206.

Danziger, L. and A. Schnytzer (1991), Implementing the Lindahl Voluntary-Exchange

Mechanism, European Journal of Political Economiy 7, 55-64.

19

Eckert, H. (2003), Negotiating Environmental Agreements: Regional or Federal

Authority? Journal of Environmental Economics and Management, 46, 1-24.

Eyckmans, J., S. Prost and E. Schokkaert (1993), E¢ ciency and Distribution in

Greenhouse Negotiations, Kyklos 46, 363-397.

Guttman, J. (1978), Understanding Collective Action: Matching Behavior, Amer-

ican Economic Review 68, 139-154.

Hoel, M. (1991). Global Environmental Problems: The E¤ects of Unilateral Action

Taken by one Country, Journal of Environmental Economics and Management

20, 55-70.

Kaul, I., Conceicao, P., Le Goulven, K. and R.U. Mendoza (2003), How to Im-

prove the Provision of Global Public Goods, in: Paul, I. et al. (eds.), Provid-

ing Global public Godds: Managing Globalization, Oxford University Press,

Oxford, UK.

Kockesen, L. and E.A. Ok (2004), Strategic delegation by unobservable incentive

contracts, Review of Economic Studies 71,397-424.

Musgrave, R.A. (1959), The Theory of Public Finance, McGraw Hill, New York.

Muthoo, A. (1996), A Bargaining Model Based on the Commitment Tactic, Journal

of Economic Theory 69, 134-152.

Muthoo, A. (1999), Bargaining Theory with Applications, Cambridge University

Press, Cambridge (UK).

Oates, W. (1972), Fiscal Federalism, New York, Harcourt-Brace-Jovanovich.

Olson, M. (1969), The Principle of Fiscal Equivalence: The Devision of Respon-

sibilities among Di¤erent Levels of Government, American Economic Review,

Papers and Proceedings 59, 479-487.

Siqueira, K. (2003), International Externalities, Strategic Interaction, and Domes-

tic Politics, Journal of Environmental Economics and Management 45, 674-

691.

Segendor¤, B. (1998), Delegation and Threat in Bargaining, Games and Economic

Behavior 23, 266-283.

Sobel, J. (1981), Distortion of Utilities and the Bargaining Problem, Econometrica

49, 597-619.

20